RESEARCH

Rift Valley Fever Outbreak with East-Central African Virus Lineage in , 2003 Ousmane Faye,* Mawlouth Diallo,* Djibril Diop,† O. Elmamy Bezeid,‡ Hampathé Bâ,† Mbayame Niang,* Ibrahima Dia,* Sid Ahmed Ould Mohamed,† Kader Ndiaye,* Diawo Diallo,* Peinda Ogo Ly,* Boubacar Diallo,‡ Pierre Nabeth,* François Simon,* Baïdy Lô,† and Ousmane Madiagne Diop*

In October 2003, 9 human cases of hemorrhagic single-stranded RNA segments referred to as L (large), M fever were reported in 3 provinces of Mauritania, West Afri- (medium), and S (small) (5). ca. Test results showed acute Rift Valley fever virus (RVFV) In West Africa, the fi rst extensive RVF outbreak record- infection, and a fi eld investigation found recent circulation of ed to date occurred in Mauritania in 1987 and resulted in 220 RVFV with a prevalence rate of 25.5% (25/98) and 4 deaths human deaths (6). After this outbreak, an active surveillance among the 25 laboratory-confi rmed case-patients. Immuno- system led to the detection of several animal cases in Mau- globulin M against RVFV was found in 46% (25/54) of do- mestic animals. RVFV was also isolated from the mosquito ritania, Senegal, and other West African countries (7–9). species Culex poicilipes. Genetic comparison of virion seg- Furthermore, during interepizootic periods, RVFV has been ments indicated little variation among the strains isolated. repeatedly isolated from different mosquito species in Sen- However, phylogenetic studies clearly demonstrated that egal, Burkina Faso, and Nigeria (10–12,). During 1998, an these strains belonged to the East-Central African lineage outbreak of RVF occurred in southeastern Mauritania, re- for all segments. To our knowledge, this is the fi rst time vi- sulting in 300 to 400 human cases and 6 deaths (13). ruses of this lineage have been observed in an outbreak in In Mauritania in 2000, health authorities established West Africa. Whether these strains were introduced or are a National Disease Surveillance System (NDSS) by using endemic in West Africa remains to be determined. sentinel herds in 5 geographic regions and a notifi cation system of hemorrhagic fever in medical healthcare centers. ift Valley fever (RVF) is an acute febrile viral disease This NDSS was implemented in collaboration with the Rthat affects small domestic ruminants (1) and humans. Centre National d’Hygiène, the “Centre National d’Elevage The disease in animals is characterized by high rates of et de Recherche Vétérinaire” in Mauritania, and the Insti- abortion and death of young ruminants (2). In humans, the tut Pasteur de Dakar in Senegal. The value of the NDSS symptoms are usually mild, but in severe cases, hemor- was further reinforced after an outbreak of Crimean-Congo rhages, meningoencephalitis, retinopathy, and sometimes hemorrhagic fever in Mauritania in February 2003 (14) for death can occur (3). The disease is widespread in Africa, which RVFV was identifi ed in animal and human serum mainly in the sub-Saharan region but also in Egypt. In specimens collected in September and October 2003. We 2000, outbreaks were recorded for the fi rst time outside of describe the results of a multidisciplinary investigation to the African continent, in Saudi Arabia and Yemen (4). RVF determine extent of the outbreak and the key factors re- virus (RVFV) belongs to the family Bunyaviridae, genus sponsible for RVFV reemergence in Mauritania. Phlebovirus genus, and its genome consists of 3 negative Materials and Methods *Institut Pasteur de Dakar, Dakar, Sénégal; †Institut National de Recherche en Santé Publique, Nouakchott, Mauritania; and ‡Cen- Case Defi nitions tre National d’Elevage et de Recherche Vétérinaire, Nouakchott, A suspected human RVFV case-patient was defi ned as Mauritania a person with fever associated or not with hemorrhagic,

1016 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 13, No. 7, July 2007 Rift Valley Fever Outbreak jaundice, or neurologic symptoms or any person who died who had had overt hemorrhagic fever symptoms from September through December 2003. A confi rmed human RVFV case-patient was defi ned as a person for whom labo- ratory tests confi rmed an acute or recent RVFV infection, e.g., by >1 of these results: immunoglobulin M (IgM), re- verse transcription–PCR (RT-PCR) or virus isolation posi- tive results. A human contact was defi ned as any person or relative who had been directly in contact with a confi rmed human or animal case-patient, or with a person who died who had had overt hemorrhagic fever symptoms from Sep- tember through December 2003. Figure 1. Locations of the study sites. Field Investigations Laboratory Tests Study Sites Nine localities belonging to 3 administrative provinces Serologic Studies were visited (Figure 1): Keur Macène and Rkiz (Trarza All human and animal samples were tested for evidence Province), Makhtar Lahjar, Guimi, Taïba, and Sagle Moure of IgG and IgM by using an ELISA technique (18,19). Se- (Brakna Province), Legrane, Kélébélé, and Hseytine (As- rum specimens were considered positive for antibodies if saba Province). These localities were chosen because they the difference between the sample and control optical den- had confi rmed human or animal cases. sities was >3 standard deviations above the mean of the negative controls. Human Investigations In affected areas, the investigation was conducted Molecular Studies under the supervision of the chief of the sanitary district. Viral RNA was extracted from serum of suspected For each case, venous or capillary blood samples were col- case-patients by using the QIAamp RNA kit (QIAGEN, lected into dry tubes or onto fi lter papers, respectively. A Inc. Chatsworth, CA, USA) and RT-PCR was done by us- thick blood smear was also taken from all suspected case- ing the Titan One-Step RT-PCR System (Roche Diagnos- patients for differential diagnosis of malaria. An interview tics, Mannheim, Germany), according to the recommenda- in which information was gathered about sex, age, date of tions of the manufacturers. The primers NS3a (nt 710–729; fever onset, and hemorrhagic signs was conducted for all 5′-ATGCTGGGAAGTGATGAGCG-3′) and NS2g (nt case-patients and their contacts. 61–80; 5′-TGATTTGCAGAGTGGTCGTC-3′) were used to amplify a 669-nt region of the virus S segment region Animal Investigations encoding the NSs protein. The primers MRV1a (nt 772– All domestic animals living in the close vicinity of 790; 5′-CAAATGACTACCAGTCAGC-3′) and MRV2g suspected or confi rmed case-patients were included in the (nt 1563–1580; 5′-GGTGGAAGGACTCTGCGA-3′) were study. Every blood sample was accompanied by an investi- used to amplify a 809-nt region of the virus M segment re- gation form specifying the species, age, and localization of gion encoding the G2 protein. Primers Wag (nt 4440–4457; the animal during the month before the investigation and, 5′-ATTCTTATTCCCGAATAT-3′) and Xg (nt 4634–4651; for female animals, a history of pregnancies. 5′-TTGTTTTGCCTATCCTAC-3′) were used to amplify a 212-nt region of the L segment (20–22). The PCR products Entomologic Investigations were purifi ed on agarose gel and directly sequenced by us- Adult mosquitoes were collected in CDC light-traps ing the Sanger method with an ABI 377 sequencer (Ap-

(15), with or without CO2, which were placed close to plied Biosystems, Foster City, CA, USA). Phylogenetic water points or in sheepfolds, respectively; animal-baited trees on the partial sequences of the S (601 nt), M (726 nt), traps were placed in the houses of persons with suspected and L (121 nt) RNA segments were constructed by using or confi rmed cases. Mosquitoes were frozen and subse- the maximum likelihood method (PAUP* 4.0, Sinauer As- quently identifi ed on a chilled table by using morphologic sociates Inc., Sunderland, MA, USA). keys (16,17). They were classifi ed into monospecifi c pools, stored in liquid nitrogen, and transported to the laboratory, Virologic Studies where they were kept at –80°C until virus isolation was Virus isolation was performed at the World Health attempted. Organization (WHO) Collaborating Center for Arbovi-

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 13, No. 7, July 2007 1017 RESEARCH ruses (www.pasteur.fr/recherche/banques/crora), Insti- tut Pasteur de Dakar on mosquitoes and serum collected from humans by inoculating the virus into suckling mice and a mosquito cell line (AP61). Virus identifi cation was performed by an indirect immunofl uorescence assay that used polyclonal and monoclonal antibodies. The identi- fi cation of virus isolates was confi rmed by complement fi xation (23).

Parasitologic Test To rule out malaria infection, thick blood smears from patients with suspected cases were Giemsa-stained. The ra- tio of parasite (Plasmodium falciparum) to leukocytes was estimated in 200 fi elds based on a mean leukocyte count of Figure 2. Investigation of human and animal contact around index 8,000/μL of blood. and suspected case-patients (S) from Mauritania in 2003. For each case-patient (represented as a box), PCR, immunoglobulin M (IgM), or isolation (Isol)-positive test results are indicated below the Results sample number (e.g., 169867). (a/b), no. IgM positive/no. tested; S*, suspected case-patient before fi eld investigation and subsequently Human Cases confi rmed positive by laboratory tests. The different cases recorded and the linkages between them are represented in Figure 2. The 9 confi rmed case- patients, identifi ed before the investigation, were from the was isolated from a blood sample taken during his illness. Assaba, Brakna, Trarza, and Tagant provinces. Subsequent No RVFV infection was noticed in contact persons associ- discussions with the head of the sanitary district enabled ated with this patient. However, IgM against RVFV was the localization of the residences and the relatives for 2 of detected in animals living in and near the residence of in- them (index case-patients 1 and 2). Several confi rmed cas- dex case-patient 1. For index case-patient 2 (infected with es-patients (index case-patients 3 to 9) and their relatives the strain SHM169872), a second blood sample was taken, were not found due to the great distances between locali- and presence of IgM antibodies against RVFV was con- ties or nomadic behaviors of populations. However, for all fi rmed. No evidence of RVFV infection was detected in suspected case-patients (S), further investigation was con- contact persons or in animals living in or near the residence ducted in the provinces where the confi rmed case-patients of this patient. The suspected case-patient 1 was identifi ed lived. in Hospital, where he was admitted on November 3, In total, 98 persons (66 contacts, 23 with suspected 2003. This 50-year-old patient had onset of fever on Octo- cases, and 9 with confi rmed cases) were included in this ber 24, 2003, with asthenia, jaundice, nausea, hematemesis, study. Of these persons, 25.5% (25/98) had evidence of re- epistaxis, and gingival hemorrhage. Serologic tests showed cent RVFV infection (i.e., presence of IgM, viral RNA or IgM against RVFV. No evidence of virus infection was virus, or >1 of these results), and 10% had evidence of past detected in those who had accompanied this patient to the infection (i.e., presence of IgG alone). Seven viral strains hospital. were isolated. For the 25 patients who were recently in- In Brakna Province, where 3 cases were confi rmed fected (9 with confi rmed cases before the investigation, 6 before the investigation (index case-patients 3–5), 6 sus- contacts, and 10 with suspected cases at the time of investi- pected case-patients (S2–S7) were found during the inves- gation), the median age was 21 years (range 7–50 years), 4 tigation. Laboratory testing of samples from the 3 index died, and 16 had hemorrhagic signs (hematemesis, vaginal case-patients showed IgM against RVFV by ELISA and bleeding, severe hemoptysis, bleeding from the gums and RVF viral RNA by RT-PCR. A virus strain (SHM169898) venipuncture sites, petechial rashes, and ecchymoses of the was isolated from index case-patient 4. Among the sus- skin). Among the 23 suspected case-patients from whom pected case-patients from this province, 3 deaths (S2–S4) blood samples were collected, 10 were infected by RVFV; were recorded, but no samples were available from those only 2 patients were positive for malaria parasites. patients. Nevertheless, animals in the vicinity of S1, S2, In Assaba Province, 2 confi rmed case-patients were re- and S3, were found to be infected (5/7, 5/8, and 3/10 ani- corded before the investigation (index case-patients 1 and mals, respectively). In addition, infection was detected in 4 2), and 1 suspected case-patient (S1) was found during the of 6 persons who had been in contact with suspected case- investigation. The index case-patient 1 was dead at the time patient 3. In contrast, no evidence of infection by RVFV of the investigation; however, RVFV (strain SHM169867) was detected in contacts of S2 and S4. The suspected case-

1018 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 13, No. 7, July 2007 Rift Valley Fever Outbreak patient 5 was a student living in Makhtar Lahjar, who had signs. The RVF IgM test result of this patient was positive. fever onset on October 20, 2003, was admitted to the Na- In Tagant Province, index case-patient 8, whose condition tional Hospital Center (NHC) of Nouakchott on November was diagnosed before the investigation, was a man who 6, and from whose blood RVFV (strain SHM172776) was came for consultation at the provincial hospital on Septem- subsequently isolated. During the investigation, 2 suspected ber 24, exhibiting fever and hematemesis. RVFV (strain case-patients (S6 and S7) were discovered in the Healthcare SHM169868) was isolated from a blood sample taken on Center of Makhtar Lahjar. Suspected case-patient 6 was a September 29. 17-year-old female patient who had onset of fever on Oc- In Gorgol Province, 2 suspected case-patients (S12 and tober 10 and who was admitted to the center on November S13) were evacuated to the NHC of Nouakchott. The onset 1 with headache, abdominal pain, vomiting, hemorrhages, of their symptoms dated to October 23 and October 25, re- and epistaxis; virus was isolated from this patient (strain spectively. Samples from each were positive for IgM against SHM172805). The contacts associated with suspected RVFV by October 30, and RVFV (strain SHM172768) was case-patient 6 were also found to be infected with RVFV. isolated from S13, who died on October 30. Suspected case-patient 7 was a 35-year-old man. He had a In Dakar, Senegal, an “imported” case (index case-pa- fever on October 15 and was admitted to the Health Center tient 9) in a person from , Mauritania, was diagnosed of Makhtar Lahjar on November 1 with headache, nausea, by positive results by ELISA IgM and RT-PCR. This pa- vomiting, and epistaxis. The RVF IgM test result for this tient was fi rst admitted to the NHC of Nouakchott, Mauri- patient was positive for RVFV. tania, before being transferred to the Hôpital Principal de In Trarza Province, 2 confi rmed case-patients were Dakar. observed during the period of surveillance (index case- patients 6 and 7) and 4 suspected case-patients (S8–S11) Animal Cases were identifi ed during the investigation. Viral isolation was Serum samples were obtained and tested from 54 do- positive for the index case-patient 7 (strain SHM169885). mestic animals (48 goats and 6 sheep) living in the visited Among the suspected case-patients, 2 (S8 and S9) died localities (Table 1). The median age was 4 years (range before blood samples could be obtained. However, blood 1–15), and the abortion rate was 70% during the last gesta- testing of samples from 2 animals living in the vicinity of tion period. IgM against RVFV was detected by ELISA in S8 and from 2 human contacts of S9 found recent RVFV 25 of 54 animals; no IgG against RVFV was found in any infection. Suspected case-patient 10 was 28-year-old man, of the 54. Among the animals with a positive test result, the with onset of fever on October 22, who came for consulta- abortion rate was 92%. tion to Keur Macene Healthcare Center. He had a prolonged cough without hemorrhagic symptoms, and an IgM ELISA Mosquitoes result for RVFV was positive. The suspected case-patient A total of 22,201 mosquitoes, belonging to 4 genera 11 is a 26 year-old woman, with onset of fever on October and 17 species, were collected. Culex poicilipes was the 22, who was admitted to the healthcare center of Rkiz with most frequent species (43.8%), followed by Cx. antenna- headaches, asthenia and anorexia without hemorrhagic tus (23%) and Mansonia uniformis (9%). A total of 544

Table 1. Prevalence rate of immunoglobulin M against Rift Valley fever virus in livestock from Mauritania, 2003 Livestock species (no. positive/no. tested) No. abortions/ District/Locality Month sampled Sheep Goats Total no. tested Brakna Taiba Oct 0/1 3/9 3/10 6/10 Guimi Oct 1/1 4/7 5/8 8/8 Sagle Moure Nov 0/0 5/7 5/7 5/7 13/25 (42%) 19/25 (76%) Trarza Boynayé Oct 0/3 0/6 0/9 0/9 Rkiz Oct 0/0 2/2 2/2 2/2 2/11 (18%) 2/11 (18%) Assaba Legrane Nov 0/1 2/5 2/6 5/6 Kélébélé Nov 0/0 1/2 1/2 2/2 Tézékré Nov 0/0 1/2 1/2 2/2 Hseytine Nov 0/0 6/8 6/8 8/8 10/18 (55%) 17/18 (94%) Total 1/6 (16%) 24/48 (50%) 25/54 (39%) 38/54 (70%)

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 13, No. 7, July 2007 1019 RESEARCH monospecifi c pools were constituted and submitted for vi- during the epidemics of RVF in Kenya (25), as well as in ral isolation. Only Cx. poicilipes was found to be associated Saudi Arabia and Yemen (4), demonstrate the importance with RVFV. Three strains (ArD 174367, ArD 174303, and of combining diagnostic assays for accurate and compre- ArD 174347) were isolated from the 146 pools constituted hensive detection of RVFV infection. in Guimi Province, giving rise to a minimum infection rate Regarding differential diagnosis, only 2 suspected of 0.04% for this locality and 0.01% for the whole study case-patients with fever had confi rmed malaria due to in- site (Table 2). fection by P. falciparum. This low malaria infection rate suggests that the RVF outbreak was the major cause of the Genetic Analysis febrile cases notifi ed during this period. RNA was extracted from the 8 viral strains isolated Although WHO estimates that the human mortality from humans, and fragments of the S, M, and L segments rate due to RVFV is ≈1%–2% of infected patients, the num- were amplifi ed and sequenced. No amino acid (aa) dif- ber of recorded deaths during this outbreak was 4 among 25 ferences were found between the fragments of the S or L infected patients when the laboratory data were considered segments analyzed (198 and 51 aa, respectively). A single exclusively. Epidemiologic investigations have found 5 ad- amino acid difference was found between the M fragment ditional deaths that could be due to RVFV infection. We (255 aa) of the 5 viral strains analyzed. Results of phylo- cannot be absolutely certain about the causes of death in genetic analyses of the nucleotide sequences of amplifi ed our suspected case-patients from whom no blood sample fragments from 3 segments belonging to 2 representa- was taken. However, when the clinical symptoms and the tive strains (H1MAU03 [SHM169867] and H2MAU03 rate of infection in domestic animals are considered, that [SHM169868]) isolated during this epidemic and previ- these cases were the result of RVFV infection is highly ously described nucleotide sequences of RVFV are shown probable. In those cases in which the contacts had negative in Figure 3. The strains identifi ed in Mauritania 2003 are test results and only domestic animals had positive results, consistently located within the East/Central lineage for all we hypothesize that the infection of those with lethal cases trees. This lineage contains viral strains that circulated in was related to socioeconomic/professional activity. Indeed, Madagascar (1991), Kenya (1997), Chad (2001), and Saudi those at highest risk include butchers and others who come Arabia (2001). in contact with animals (e.g., slaughterhouse workers, tan- ners, and herdsmen), who represent a large part of the pop- Discussion ulation living in these areas. The combination of ELISA, RT-PCR, and isolation During this investigation, a high infection rate was assays has permitted the rapid and effi cient identifi cation found in sheep and goats that lived in close proximity to of RVFV as the cause of the extended hemorrhagic fever the patients (46.3% of IgM positive compared with 25% outbreak reported in Mauritania during the last quarter of during the 1998 outbreak) (13). Also, according to inter- 2003. Of the 24 RVF cases diagnosed in the laboratory, 13 views with herdsmen, a high abortion rate (92%) was ob- were diagnosed by IgM only; 8 were diagnosed by IgM, served in infected animals during their most recent preg- RT-PCR, isolation, or >1 method; and 3 were diagnosed by nancy. Previous studies showed abortion rates ranging RT-PCR, isolation, or both. These data and those obtained from 80% to 100% (26). Nevertheless, the discrepancies

Table 2. Mosquitoes collected during the Rift Valley fever outbreak, Mauritania, 2003 Culex Cx. Mansomia antennatus Cx. poicilipes tritaeniorhynchus uniformis Others* Total Locality CPCPC P CP C P C P K. Macene 273 8 420 12 195 5 1,755 67 365 18 3,008 110 Boynayé 46 1 137 4 766 17 159 5 667 14 1,775 41 Techtayatt 696 14 290 7 451 9 30 1 627 15 2,094 46 1,694 35 1437 30 155 3 69 2 435 12 3,790 82 Taiba 2,243 45 255 7 38 2 540 17 3,076 71 Boghe 16 3 16 3 Guimi 159 3 7163 146† 943 21 8,265 170 Legrane 6 1 14 2 20 3 Kélébélé 11 2 23 4 34 6 Hseytine 83 8 3 Sarandougu 8 1 18 1 3 1 86 6 115 9 Total 5,119 107 9,737 210 1,608 37 2,013 75 3,724 87 22,201 544 *Aedes. vexans, Ae. ochraceus, Anopheles funestus, An. gambiae, An. pharoensis, An. rufipes, An. wellcomei, An. ziemanni, An. squamosus, Cx. ethiopicus,Cx. neavei and Ma. African. C, collected; P, pools. †Rift Valley fever virus strains (ArD 174367, ArD 174303, and ArD 174347) isolated in 3 pools.

1020 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 13, No. 7, July 2007 Rift Valley Fever Outbreak

Figure 3. Phylogenetic relationships of the S (small), M (medium), and L (large) RNA segments of Rift Valley fever viruses. Strains isolated in Mauritania (gray shading) are designated H1MAU03 and H2MAU03, according to previous abbreviation guidelines (24). Nucleotide sequences of these segments (S, M, and L) have been submitted to GenBank with the following accession nos., respectively: EF160113, EF160116, and EF160117 for H1MAU03; EF160114, EF160115, and EF160118 for H2MAU03. Branch lengths are proportional to the number of substitutions per site. observed between the overall abortion rate in animals and Genetic analyses of the 3 segments of RVFV isolated the prevalence of anti-RVF IgM (70% vs 46.3%) support during this epidemic showed a low level of variation be- the hypothesis of the existence of other cocirculating dis- tween isolates from the different provinces. This fi nding eases that also cause abortion. The lack of anti-RVF IgG in supports the hypothesis that the same strain was circulating domestic animals is surprising considering the long-stand- in the different affected areas. The nucleotide sequences ing virus endemicity in Mauritania. This observation could of the strains isolated during this epidemic compared be ascribed to different factors: 1) the small number of do- with those isolated elsewhere in Africa and Saudi Arabia mestic animals analyzed, 2) a recent introduction of the vi- showed that they belong to the East/Central African cluster rus in these localities, or 3) a renewal of animal populations for the 3 segments. Previous reports have shown that some (27). The latter hypothesis is supported by the observation strains isolated in West Africa share 1 or 2 segments with of the relatively young median age of the animals tested strains belonging to the East/Central African cluster (2,29). (4 years old), refl ecting a new animal population since the However, to our knowledge, this is the fi rst evidence of the 1998 RVF epidemic in Mauritania. circulation in West Africa of strains harboring 3 segments Among the mosquitoes collected, several species that all belong to the East/Central African cluster. known as RVFV vectors (Cx. poicilipes, Cx. tritaenio- This fi nding confi rms the existence of RVFV strain rhynchus, Cx. antennatus, An. pharoensis, Ae. vexans, exchanges between geographic areas. In fact, the spread of Ae. ochraceus, and Ma. africana) (10,28) were recorded, RVFV from East Africa to other regions has already been but only Cx. poicilipes was found to carry RVFV dur- observed during the RVF outbreak in Saudi Arabia and Ye- ing the outbreak. Spatial analyses of the results show that men in 2000–2001 (4) and in Chad in 2001 (30). RVFV Cx. poicilipes was in fact predominant only in the village was also found to be the cause of the epidemic/epizootic of Guimi, where the RVFV was isolated. This observa- in Egypt in 1977 and in Madagascar in 1979 (31). Such tion indicates that the levels of the different species vary a mechanism of RVFV spread likely depends on human according to the local environment. Mosquitoes from the and animal population movements for which animal migra- Aedes genus, known for their role in RVFV maintenance tion routes between West and East/Central Africa need to and transmission, were scarce, likely due to their early be identifi ed. Furthermore, previous studies have demon- appearance at the onset of the rainy season, whereas our strated that reassortant viruses can emerge when 2 RVFV investigation took place at the end of the rainy season. lineages coexist (24). Favorable environmental conditions Indeed, in 2003, the last rainfall event was recorded at the (mainly the rainfall pattern), which led to the emergence beginning of October. of already introduced East/Central African strains, seem to

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 13, No. 7, July 2007 1021 RESEARCH be the cause of this outbreak. Indeed, RVFV emergence 7. Zeller HG, Ba MM, Akakpo JA. Rift Valley fever epizootic in small in East Africa is undoubtedly the consequence of rainfall ruminants in southern Mauritania (October 1993): risk of extensive outbreaks. Ann Soc Belg Med Trop. 1995;75:135–40. surplus (32,33). In contrast, in West Africa, the few studies 8. Zeller HG, Fontenille D, Traore-Lamizana M, Thiongane Y, Di- carried out in the past indicate that rainfall surplus is not a goutte JP. Enzootic activity of Rift Valley fever virus in Senegal. Am key factor for RVFV emergence. In fact, RVF outbreaks J Trop Med Hyg. 1997;56:265–72. were often observed during years of rainfall defi cit (10,34). 9. Thonnon J, Picquet M, Thiongane Y, Lo M, Sylla R, Vercruysse J. Rift Valley fever surveillance in the lower Senegal River basin: Therefore, this outbreak in Mauritania, caused by RVFV update 10 years after the epidemic. Trop Med Int Health. 1999;4: strains of the East/Central African lineage, is likely linked 580–5. to the heavy rainfall recorded during 2003 in the affected 10. Fontenille D, Traoré-Lamizana M, Diallo M, Thonnon J, Digoutte areas (315 mm in 2003 versus 161 mm in 2002). These ar- JP, Zeller H. New vector of Rift Valley fever in West Africa. Emerg Infect Dis. 1998;4:289–93. guments support the hypothesis that episodes of heavy rain- 11. Saluzzo JF, Chartier C, Bada R, Martinez D, Digoutte JP. Rift Valley falls are directly or indirectly more favorable to the emer- fever in Western Africa. Rev Elev Med Vet Pays Trop. 1987;40:215– gence of virus strains belonging to the East/Central African 23. cluster. However, such a hypothesis presumes the existence 12. Lee H. Isolation of viruses from fi eld population of Culicoides (Dip- tera Ceratopogonidae) in Nigeria. J Med Entomol. 1979;16:796–9. of ecologic or biologic differences between strains of the 2 13. Nabeth P, Kane Y, Abdalahi MO, Diallo M, Ndiaye K, Ba K, et al. lineages, and further investigations are needed with special Rift Valley fever outbreak in Mauritania in 1998: seroepidemiologic, emphasis on the interactions with the strains’ respective virologic, entomologic, and zoologic investigation. Emerg Infect vectors and reservoirs. Dis. 2001;7:1052–4. 14. Nabeth P, Cheikh DA, Lo B, Faye O, Mohamed IO, Niang M, et al. Crimean-Congo hemorrhagic fever, Mauritania. Emerg Infect Dis. Acknowledgments 2004;10:2143–9. We are indebted to Roughiètou Sylla, Maguèye Ndiaye, Jo- 15. Sudia WD, Chamberlain RW. Battery-operated light trap, an im- seph Faye, Lang Girault, Mireille Mondo, Modou Diagne, and proved model. [originally published 1962]. J Am Mosq Control As- soc. 1988;4:536–8. Mamadou Diallo for their excellent technical assistance in labora- 16. Edwards FW. Mosquitoes of the Ethiopian region. III Culicine adult tory diagnosis as well as fi eld investigations. We thank the au- and pupae. London: British Museum (Natural History); 1994. thorities and the fi eld agents of the Ministry of Health and the 17. Service MW. Handbook to the Afrotropical toxorhynchytine and Ministry of Livestock in Mauritania for facilitating the investiga- culicine mosquitoes excepting Aedes and Culex. London: British Museum (Natural History); 1993. tion of this outbreak. We are grateful to Mike McCune for gram- 18. Niklasson B, Peters CJ, Grandien M, Wood O. Detection of hu- matical editing. man immunoglobulins G and M antibodies to Rift Valley fever virus by enzyme-linked immunosorbent assay. J Clin Microbiol. This work was supported by grants from the Institut Pasteur 1984;19:225–9. de Dakar, Senegal. 19. Monath TP, Nystrom RR. Detection of yellow fever virus in serum by enzyme immunoassay. Am J Trop Med Hyg. 1984;33:151–7. Dr Faye is a virologist and acarology specialist who works 20. Muller R, Saluzzo JF, Lopez N, Dreier T, Turell M, Smith J, et al. at the Department of Virology, Institut Pasteur de Dakar, Senegal. Characterization of clone 13, a naturally attenuated avirulent isolate His primary research interests include the virology and vectorial of Rift Valley fever virus, which is altered in the small segment. Am J Trop Med Hyg. 1995;53:405–11. transmission of arboviruses and viral hemorrhagic fevers. 21. Takehara K, Min MK, Battles JK, Sugiyama K, Emery VC, Dalrym- ple JM, et al. Identifi cation of mutations in the M RNA of a candidate vaccine strain of Rift Valley fever virus. Virology. 1989;169:452–7. References 22. Muller R, Poch O, Delarue M, Bishop DHL, Bouloy M. Rift Valley 1. Daubney R, Hudson JR, Garnham PC. Enzootic hepatitis or Rift Val- fever virus L segment: correction of the sequence and possible func- ley fever: an undescribed virus disease of sheep, cattle, and man tional role of newly identifi ed regions conserved in RNA-dependent from east Africa. J Pathol Bacteriol. 1931;34:545–79. polymerases. J Gen Virol. 1994;75:1345–52. 2. Easterday BC, McGavran MH, Rooney JR, Murphy LC. The 23. Digoutte JP, Calvo-Wilson MA, Mondo M, Traore-Lamizana M, pathogenesis of Rift Valley fever in lambs. Am J Vet Res. 1962;23: Adam F. Continuous cell lines and immune ascitic fl uid pools in 470–9. arbovirus detection. Res Virol. 1992;143:417–22. 3. Laughlin LW, Meegan JM, Strausbaugh LJ, Morens DM, Watten 24. Sall AA, de Zanotto PM, Sene OK, Zeller HG, Digoutte JP, Thi- RH. Epidemic Rift Valley fever in Egypt: observations of the spec- ongane Y, et al. Genetic reassortment of Rift Valley fever virus in trum of human illness. Trans R Soc Trop Med Hyg. 1979;73:630–3. nature. J Virol. 1999;73:8196–200. 4. Shoemaker T, Boulianne C, Vincent MJ, Pezzanite L, Al-Qahtani 25. Woods CW, Karpati AM, Grein T, McCarthy N, Gaturuku P, Muchiri MM, Al-Mazrou Y, et al. Genetic analysis of viruses associated with E, et al. An Outbreak of Rift Valley fever in Northeastern Kenya, emergence of Rift Valley fever in Saudi Arabia and Yemen, 2000– 1997–98. Emerg Infect Dis. 2002;8:138–44. 01. Emerg Infect Dis. 2002;8:1415–20. 26. Coetzer JA. The pathology of Rift Valley fever. II. Lesions occurring 5. Schmaljohn CS, Beaty BJ, Calisher CH, et al. Family Bunyaviridae. in fi eld cases in adult cattle, calves and aborted fetuses. Onderste- Virus taxonomy: classifi cation and nomenclature of viruses. Sixth poort J Vet Res. 1982;49:11–7. report of the International Committee on Taxonomy of Viruses. Arch 27. Thiongane Y, Thonnon J, Zeller H, Lo MM, Faty A, Diagne F, et Virol Suppl. 1995;10:300–15. al. Recent data on Rift Valley fever epidemiology in Senegal [in 6. Digoutte JP, Peters CJ. General aspects of the 1987 Rift Valley fever French]. Dakar Med. 1996; Special issue:1–6. epidemic in Mauritania. Res Virol. 1989;140:27–30.

1022 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 13, No. 7, July 2007 Rift Valley Fever Outbreak

28. Diallo M, Nabeth P, Ba K, Sall AA, Ba Y, Mondo M, et al. Mos- 32. Davies FG, Linthicum KJ, James AD. Rainfall and epizootic Rift quito vectors of the 1998–1999 Rift Valley fever (RVF) outbreak, Valley fever. Bull World Health Organ. 1985;63:941–3. and others arboviruses (Bagaza, Sanar, West Nile and Wesselsbron), 33. Linthicum KJ, Anyamba A, Tucker CJ, Kelley PW, Myers MF, Pe- in Mauritania and Senegal. Med Vet Entomol. 2005;19:119–26. ters CJ. Climate and satellite indicators to forecast Rift Valley fever 29. Centers for Disease Control and Prevention. Outbreak of Rift Valley epidemics in Kenya. Science. 1999;285:397–400. fever, Saudi Arabia, August–November 2000. MMWR Morb Mortal 34. Ndione JA, Kébé CMF. Epizootie de la fi èvre de la Vallée du Rift en Wkly Rep. 2000;49:982–5. 1987: conditions pluviométriques d’une émergence. Pub Assoc Int. 30. Durand JP, Bouloy M, Richecoeur L, Peyrefi tte CN, Tolou H. Rift 2002;14:428–36. Valley fever virus infection among French troops in Chad. Emerg Infect Dis. 2003;9:751–2. Address for correspondence: Ousmane Madiagne Diop, Département de 31. Sall AA, Zanotto PM, Vialat P, Sene OK, Bouloy M. Molecular epi- Virologie, Institut Pasteur de Dakar, 36 Ave Pasteur, Boîte Postale 220, demiology and emergence of Rift Valley fever. Mem Inst Oswaldo Cruz. 1998;93:609–14. Dakar, Sénégal; email: [email protected]

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 13, No. 7, July 2007 1023